cd /news/ai-safety/oxide-joins-anthropic-s-project-glas… · home topics ai-safety article
[ARTICLE · art-76890] src=oxide.computer ↗ pub= topic=ai-safety verified=true sentiment=↑ positive

Oxide joins Anthropic's Project Glasswing

Oxide Computer Company has joined Anthropic's Project Glasswing, a collaborative effort to secure critical software, and is applying Anthropic's Claude Mythos 5 to proactively identify and patch vulnerabilities across its full software stack, from firmware to the network stack. The company, which designs transparent and verifiable infrastructure, will share its security work openly, building on its earlier discovery of a gap in the NXP LPC55S69's TrustZone.

read2 min views1 publishedJul 28, 2026
Oxide joins Anthropic's Project Glasswing
Image: Oxide (auto-discovered)

Joining a collaborative effort to secure the world’s most critical software #

Everyone who has operated their own compute infrastructure has experienced the challenge of auditing firmware, microcode, embedded controllers, and operating systems — all software that are prime targets for bad actors. That experience shaped our belief that all critical software should be inspectable, and is why we open source everything we write. Today we’re joining Anthropic’s Project Glasswing, a collaborative effort to secure the world’s most important software. As part of this collaboration, Oxide is applying Claude Mythos 5 to proactively identify and patch potential vulnerabilities in our own codebase — from firmware through the operating system, control plane, and network stack.

Our core engineering belief at Oxide is that hardware and software should be co-designed — we control every layer, from the root of trust through firmware, operating system, and control plane, out to the network. This is what makes Mythos so useful to us: because we’ve designed the full stack from first principles, we can point it at every layer and act quickly on what it finds.

Co-founder and CTO

Mythos accelerates our cybersecurity defenses without sacrificing responsibility and rigor. We embrace the responsibility to share our security work in the open, including our earlier discovery of a gap in the NXP LPC55S69’s TrustZone. In RFD 576, we describe how we use large language models to bring more engineering rigor to difficult problems and sharpen our own thinking.

We built Oxide having spent years operating infrastructure that was opaque and full of places for vulnerabilities to hide. Designing transparent and verifiable infrastructure was foundational for us. That philosophy extends to how we use Mythos to pressure-test our own stack. Doing this work in the open will benefit every organization that depends on critical software.

Co-founder and CEO

To learn more about Oxide, visit [https://oxide.computer](https://oxide.computer)

To apply to join Oxide, visit [https://oxide.computer/careers](https://oxide.computer/careers)
── more in #ai-safety 4 stories · sorted by recency
── more on @oxide computer company 3 stories trending now
sponsored brought to you by zahid.host 4,200+ EU-deployed projects
reading about agents? ship yours in a single git push.

Run your AI side-project on zahid.host

EU-based hosting, git-push deploys, automatic HTTPS, no cold starts. Free tier with a custom domain — perfect for shipping the agent you just read about.

$git push zahid main
Live at https://your-agent.zahid.host
Get free account → Pricing
from €0/mo · no card required
LIVE [news/oxide-joins-anthropi…] indexed:0 read:2min 2026-07-28 ·